To Tell Or Not To Tell?Assessing the Practical Effects of Disclosure for Word-of-Mouth Marketing Agents and Their Conversational Partners

A Summary Report1

January 2006

Walter J. Carl, Ph.D.Assistant Professor Department of Communication Studies Northeastern University

2

Table of Contents

Table of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Defining Key Terms Used In This Report With the WOMMA Terminology Framework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

Disclosure Policy at Time of Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

What We Learned About Business Outcome Metrics . . . . . . . . . . . . . . . . . . . . . . . . 8

WOM Credibility Effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

Inquiry, Use, Purchase, and Pass-Along/Relay Outcomes. . . . . . . . . . . . . . . . 9

Response Rate According To Relationship Type. . . . . . . . . . . . . . . . . . . . . . . 9

How Business Outcome Metrics Were Affected By Agent Disclosure . . . . . . . . . . . 10

How Many Agents Disclosed Their Identity? . . . . . . . . . . . . . . . . . . . . . . . 10

Are Certain Types of Relationships More Likely To Know Agent’s Identity Rather Than Others? . . . . . . . . . . . . . . . . . . . . . . . . 11

How Did Conversational Partner Learn of Agent’s Identity? . . . . . . . . . . . . 11

What Type of Information Did The Agents Share? . . . . . . . . . . . . . . . . . . . 12

What Are Conversational Partners’ Feelings Towards Agent Affiliation With The Marketing Company?. . . . . . . . . . . . . . . . . . . . 12

Does Knowing Agent Affiliation Matter To Outcomes?. . . . . . . . . . . . . . . . 13

Conclusion and Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

Contact Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

Copyright and Citation Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

Appendix A – Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

Appendix B – Methodology and Limitations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

Notes and References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

3

SummaryIncreasingly, companies have sought to harness the power of word-of-mouth com-munication by enlisting consumers to talk about brands, products, and services within their peer networks. Whether those consumers are paid by, or voluntarily affiliate with, a marketing organization, or are part of the company’s existing customer database, ethical concerns about their word-of-mouth practices have been raised by various social critics, especially in terms of whether or not the consumers disclose their partici-pation in the organized program/campaign when talking with others. Further, various marketing agencies have different visions of whether or not disclosure represents an ethical imperative or if it even makes good business sense.

To test whether or not disclosing corporate affiliation has any practical business ad-vantages, Northeastern University partnered with BzzAgent, Inc., a leading word-of-mouth marketing organization, to better understand the role of disclosure in everyday and campaign-related word-of-mouth communication episodes. A novel dyadic design was employed to capture multi-party perspectives on the same episodes. Specific questions to which we sought answers included how many people affiliated with the marketing organization actually disclosed their affiliation, how they did so, what infor-mation about the marketing organization they shared, and whether or not the disclo-sure led to differences in the following key outcome measures: the credibility effect of word-of-mouth, as well as inquiry, use, purchase, and pass-along/relay intentions and behaviors.

Key findings include:

• For approximately 75% of the conversational partners (the people with whom the word-of-mouth marketing agents engaged in word-of-mouth communica-tion) it did not matter that they were talking with someone affiliated with a marketing organization. Instead what mattered was that they trusted the agent was providing an honest opinion, felt the agent had their best interests at heart, and were providing relevant and valuable information.

• None of the key outcome metrics (credibility, inquiry, use, purchase, and pass-along/relay) were negatively affected by the agent disclosing their affili-ation. In fact, the pass-along/relay rate (the number of people a person told after speaking with a word-of-mouth marketing agent) actually increased when the conversational partner was aware they were talking with a participant in an organized word-of-mouth marketing program.

• In over 75% of the cases where a person learned about a brand or product from another source of information (such as a print, radio, TV, or web adver-tisement), talking with the marketing agent increased the believability of that other source of information. This finding was also unaffected by agent disclo-sure.

• Prior to the enforcement of the word-of-mouth marketing organization’s disclosure policy (where agents were required to disclose their affiliation in epi-sodes involving an organized word-of-mouth campaign), 37% of the conversa-tional partners reported they did not know of the agent’s affiliation.

• Of seven different categories of relationships (strangers, acquaintances, friends, best friends, romantic partners/spouses, relatives, and co-workers), romantic partners/spouses, best friends, relatives, and friends were the most likely to

4

know about the agent’s affiliation, with strangers, acquaintances, and cowork-ers the least likely to know.

• For about 5% of the conversational partners who were not aware of the agent’s affiliation with the marketing organization there was a negative “backlash” effect when they found out. These negative feelings could be directed toward the agent, the interaction with that agent, the brand being discussed, and/or the company who made the brand, product, or service. There were virtually no negative feelings, however, when the conversational partner was aware of the agent’s affiliation.

• When affiliation with the marketing organization was disclosed, the agent vol-unteered the information directly to their conversational partner without being prompted 75% of the time.

• Conversational partners were most likely to know that the word-of-mouth marketing agents received free samples and that they shared their opinions with others. While nearly 80% knew that agents reported back to the market-ing organization about the WOM episode only 45% knew that the marketing organization compiled those report for a client company as part of market research.

Key conclusions include:

• Participation in an organized word-of-mouth marketing program does not undermine the effectiveness of word-of-mouth communication.

• Disclosure has practical business benefits. It does not interrupt the “natural” flow of conversation.

• Word-of-mouth marketing organizations should adopt a clear policy that requires disclosure. This policy should be implemented with a combination of both education about the practical business benefits of disclosure as well as enforcement procedures.

• Word-of-mouth marketing organizations should pay special attention to inter-actions with strangers and acquaintances as these relationship types were the least likely to know about agent affiliation and also more likely to have nega-tive feelings when they did not know about agent affiliation.

• Policies regarding disclosure should go beyond requiring agents to disclose affiliation and should have special considerations to make clear the market research aspect of the business model.

Agencies and companies who enlist the support of consumers to spread word-of-mouth can revisit their business practices and disclosure policies in light of the findings and conclusions of this report.

5

IntroductionWord-of-mouth (WOM) communication is widely seen to be a powerful force in the marketplace2. Whether it is through online or offline media, information is filtered by trusted peers thus saving time and reducing risk and uncertainty in making decisions relevant to people’s everyday lives. Further, peers are seen to be free, unbiased sources of advice because the information is given from a source that is not selling anything or tied to a corporate entity. But what happens when a recommendation is provided by a source affiliated with a corporate entity, such as a word-of-mouth marketing organiza-tion?

Controversy surrounding the ethics of certain kinds of marketing practices has grown over the last few years as the use of stealth and undercover marketing tactics have at-tracted public attention and critique3. Stealth tactics are involved anytime a person has a connection to a corporate entity, either as a paid employee or receiving some form of compensation, but without disclosing that affiliation to the people with whom they communicate. Stealth can take the form of paid actors pretending to be an “everyday person” in talking favorably about a brand and involving other consumers with that brand, bloggers who receive compensation to talk up a product without disclosing that compensation, paying or otherwise compensating people to flood chat rooms or online review sites with positive commentary about one’s own products or negative com-mentary about other company’s products, etc. Although the Word-of-Mouth Market-ing Association (WOMMA) considers stealth marketing to be unethical4, undercover tactics are seen by its proponents to be effective because, when done “well,” they are seamlessly integrated within the natural flow of everyday life5. Disclosing affiliation has often been resisted by agencies that employ stealth marketing because it is seen to un-dermine the effectiveness of the tactic. Relatedly, disclosure has been resisted because it is seen as awkward and would interrupt the natural flow of everyday interaction.

Disclosure as it applies to word-of-mouth marketing has rightfully been discussed by consumer protection, family groups, and social critics as an issue of ethics. But what are the business consequences of disclosure? Are business models effective when disclosure is required and can disclosure be done in a way that maintains the integrity of everyday personal and social relationships? In partnership with a leading word-of-mouth marketing organization, BzzAgent, Inc.6, we set out to determine the practical effects of disclosure as they relate to the following issues:

• Outcomes of WOM episodes, such as credibility of information provided, likelihood to inquire further about a product or service, likelihood of use, like-lihood of purchase, and pass-along/relay (how many other people the conversa-tional partners told as a result of interacting with a word-of-mouth marketing agent)7;

• Percentage of agents who disclosed their affiliation, how they disclosed, and what information about the marketing organization and business model was shared;

• Feelings about the conversational partner, episode, product/service, and com-pany who makes the product/service when the agent’s identity is known or not known.

We designed our study to capture dyadic, or two person, perceptions for an interac-tion; those of the person affiliated with the word-of-mouth marketing organization (or “agent”) and the person (or people) with whom the agent communicated (the “conver-

6

sational partner” or CP for short; further explanation for these terms provided below). Capturing both parties’ perceptions is essential in such a study in order to have a way to verify the accuracy of each person’s reports (for example, while agents may say they are disclosing their identity, we can feel more confident if their conversational partner also confirms that information).

Just over 800 agents and conversational partners participated in the study (211 were conversational partners). Participants were predominantly female (approximately 80%), ranged 13-79 years of age (median age of 30), and geographically distributed across the United States. The participants generated over 1,000 valid surveys that asked them to provide information on when and where the episode occurred, the topic of the conversation, whether or not a recommendation was made and by whom, the valence of the talk (positive, negative, or neutral), as well as perceptions of the quality of the interaction, the person with whom they were speaking, and how long they knew one another. Some of these brand-related conversations, or episodes, were part of a current word-of-mouth marketing campaign (which we called “Type 1” or institutional WOM episodes) and others were about products unrelated to a market-ing campaign (which we called “Type 2” or everyday WOM episodes). Throughout this report comparisons will be made between campaign and non-campaign-related WOM episodes, instances where the CP did and did not know they were talking to a person affiliated with a marketing organization, and among different relationship types (strangers versus acquaintance versus best friends, etc.). Additional details about the study’s methodology and limitations can be found in Appendix B.

Agents and CPs were able to categorize their relationship in one of the following seven ways: stranger, acquaintance, friend, best friend, romantic partner or spouse, rela-tive, or coworker. The majority of the relationships between the agents and the CPs were friends (31.8%), relatives (25.1%), romantic partners or spouses (13.7%), and best friends (12.8%), with smaller percentages represented by acquaintances (5.2%), co-workers (5.2%; though about 18% were “blended” relationships, such as both a co-worker and a friend), strangers (4.3%), and “other” (1.9%; such as a business client). The median length of time they knew one another was six years. About 15% of the CPs knew the agent less than a year and about 6% knew the agent less than a month.

Defining Key Terms Used In This Report With the WOMMA Terminology Framework

There are as many different definitions for word-of-mouth and word-of-mouth mar-keting as there are marketers and marketing companies. To address this issue, the Word of Mouth Marketing Association has developed a Terminology Framework to create non-proprietary, industry-standard terms8. This section will define key terms used in the report and will note any departures from the WOMMA Terminology Framework. The Terminology Framework defines word-of-mouth as “the act of a consumer creat-ing and/or distributing marketing-relevant information to another consumer.” More specifically, word-of-mouth is informal, evaluative (positive or negative) talk between two or more people, online or offline, about an organization, brand, product, or ser-vice, which may or may not include a recommendation. A word-of-mouth episode refers to the interaction, or portion of an interaction, that involves such marketing-related talk. Word-of-mouth can be broken down into two main types: everyday (or “organic”) and institutional (or “amplified”)9. Everyday word-of-mouth occurs when there is no corporate affiliation relevant for any of the people involved and the brand being discussed is not part of an active marketing campaign. In contrast, institutional

7

word-of-mouth involves any attempt by a company to facilitate, manage, and measure the word-of-mouth activity of people participating in an organized word-of-mouth marketing campaign or program. The distinction between institutional and everyday word-of-mouth parallels the distinction introduced above between Type 1 (campaign-related) and Type 2 (non-campaign-related) episodes.

Consistent with the Word of Mouth Marketing Association’s (WOMMA) website, this report defines word-of-mouth marketing as an umbrella term that covers a range of different marketing techniques “geared toward encouraging and helping people to talk to each other about products and services.”10 Some of these techniques include organizing and motivating volunteers to reach out to people in their social networks (grassroots marketing), using the media or celebrity endorsers to create “buzz” about a brand (buzz marketing), providing samples of products for consumers who are likely to influence the opinions of others (product seeding), and many others.

The word-of-mouth marketing organization involved in this study combines a variety of techniques in developing word-of-mouth marketing programs. Below is a descrip-tion of their business model as explained to conversational partners participating in this research project11:

The particular word-of-mouth marketing agency with which your conversa-tional partner is affiliated works with a group of volunteers (called agents) who receive a free sample of a product (or an opportunity to obtain the product at a reduced price) and information about that product. The agents use the prod-uct and share their opinions with others and then report their own and others’ experiences with the product back to the agency. In turn, the agency compiles a summary report of the feedback for the company that makes the product. Agents receive points for their reports, which they can do nothing with, re-deem for prizes, or donate to charity.

The Terminology Framework uses the generic term “Participants” to refer to the peo-ple involved in a word-of-mouth episode but does not distinguish between people who participate in an organized word-of-mouth marketing campaign and those that do not. Various word-of-mouth marketing companies use distinct terms – agents, panel-ists, seeders, influencers, etc. – to refer to the people who are either paid or voluntarily affiliate with the marketing organization (the latter are often rewarded with non-mon-etary forms of compensation such as prizes or swag). Thus, in the absence of non-pro-prietary, industry-standard alternatives to capture the distinction we needed regarding institutional affiliation (or lack thereof ) we used the term “agent” as a generic term for a person affiliated with a marketing agency and whose affiliation involves them spread-ing commentary and recommendations about a company’s brand, product, or service and “conversational partner” (or CP) to refer to people with whom agents engage in word-of-mouth episodes. We use these same terms across campaign-related (Type 1) and non-campaign-related (Type 2) episodes, with the caveat that the agent’s institu-tional identity may or may not be relevant when engaging in everyday word-of-mouth communication.

This report defines disclosure from a “receiver-oriented” view of communication, meaning we considered an agent to have disclosed when their conversational partners knew of their affiliation to the word-of-mouth marketing organization through some action taken by the agent. This definition encompasses a range of ways that the CP could learn of the agent’s identity, including the agent explicitly mentioning their af-filiation without being prompted, the CP asking the agent and then the agent telling

8

the CP, having the issue mutually come up, having the CP figure it out for themselves, or via other means (as described below).

The Terminology Framework also defines a number of outcomes. Credibility is a characteristic of participants (both agents and CPs) and assesses the believability of a message source. Consistent with communication-based measures of source cred-ibility, this report breaks credibility down into three distinct factors12: trustworthiness (defined here as being perceived as genuine and ethical), caring/goodwill (the extent to which a person feels the other has their best interest at heart), and competence (knowledge and expertise). Inquiries involve one party seeking additional information after the WOM episode, such as going to a website to learn more. Pass-along refers to how many other people the conversational partner told as a result of interacting with the agent (similar to “Relays” in the Terminology Framework). We also distinguished between purchase and use outcome metrics so that we could assess whether the CP experienced using the product or service or went on to make a purchase as well.

Finally, because the content or topic of the word-of-mouth episode could be an orga-nization/company, brand, product, or service, we will use the designation “OBPS” as a general term when reporting results, or the generic term “object” of the WOM episode. Definitions for each bold-faced term are also included in the Glossary (see Appendix A).

Disclosure Policy at Time of Study

When BzzAgent, Inc. was founded in 2001 its policy was to allow agents to make their own decision about whether or not they were going to disclose their affiliation with the marketing organization. This policy was consistent with prevailing industry views regarding the relationship between transparency and campaign effectiveness; that is, disclosing affiliation was seen to undermine the effectiveness of the communication between the person affiliated with the organized word-of-mouth program and their conversational partners. In November 2004, however, the marketing organization changed its policy and Code of Conduct to require agents to disclose a) their affiliation and b) their participation in an active campaign. Disclosing additional information, such as the specifics of how a campaign works, was left to the individual agent’s dis-cretion (that is, they were not required to disclose, or withhold, this information). As will be discussed below, the marketing organization has subsequently strengthened its disclosure policy, in part as a result of this study’s findings. Additional details about the disclosure policy are provided in the Notes section at the end of this document13.

The remainder of this report will discuss the results of the study followed by its con-clusions and recommendations. The results will be organized by what we learned about key metrics relevant to word-of-mouth marketing organizations and then how these outcomes were, or were not, affected by disclosure.

What We Learned About Business Outcome MetricsIn general, communicating with word-of-mouth marketing agents led to consequential outcomes for their conversational partners. Five of these will be discussed: credibility effect, likelihood to inquire further about the OBPS, purchase likelihood, pass-along/relay outcomes, and response rate by relationship type.

WOM Credibility Effect

About 15% of the CPs had no familiarity with the OBPS before talking with the

9

word-of-mouth marketing agent (consistent across campaign-related and non-cam-paign-related episodes). For the majority of the episodes, however, the CP had heard about the product/service from another person or a media source (such as a print, web, radio, or TV advertisement or story). In over 75% of the cases, talking with the agent increased the believability of the other source of information. In about 5% of the cases talking with the agent made the CP more skeptical of the other information source (for example, if the agent’s commentary on the product/service contradicted what the CP heard from another person or advertisement). No change was reported for 20% of the episodes. There were no significant differences between campaign- or non-campai-gn-related episodes.

Inquiry, Use, Purchase, and Pass-Along/Relay Outcomes

We asked conversational partners to indicate the likelihood that they would make a further inquiry about the organization, brand, product, or service, use it or purchase it (if relevant), or pass-along/relay information about the OBPS as a result of their conversation with the word-of-mouth marketing agent, as well as actions they had already taken (for example, how many people they had already told). The scale for the likelihood variables was 0 (not likely at all) – 9 (highly likely).

On average, CPs indicated they were fairly likely to seek out additional information (average = 6.31), with 49% indicating that they had already made an inquiry (such as visiting a website or other steps to learn more information).

CPs indicated an even greater likelihood of purchasing the brand, product, or service discussed (average = 6.96), with 40% already indicating they had purchased the pro-duct. For use of the product or service, CPs indicated a fairly high likelihood (average = 7.01), with 52% indicating they had already used the product/service.

CPs were also fairly likely to tell others about the product (average = 6.58) and told at least one other person, on average (mean = 1.65; median = 1.0). Just over 56% told at least one other person.

There were no differences between campaign (Type 1) and everyday (Type 2) organiza-tions, brands, products, or services.

Response Rate According To Relationship Type

The response rate for valid and complete surveys was 25.3%. We used the response rate for this survey as an additional metric of WOM effectiveness, akin to inquiry behavior. While agreeing to participate in a research project that requires a conversa-tional partner to visit a website and complete a survey that requires 15-20 minutes of

In over 75% of the cases where a person learned about a brand or product from another source of information (such as a print, radio, TV, or web advertisement), talking with the word-of-mouth marketing agent increased the believability of that other source of information.

The response rate for survey completion by the conversational partners, which was used as an additional metric for WOM effectiveness, was just over 25%.

10

their time as part of an academic study is admittedly different than going to a website to learn more information about a brand, the underlying dynamics of word-of-mouth influence are arguably similar.

We then matched the relationship types indicated on the agents’ surveys with those indicated on the surveys completed by CPs to determine the response rate based on each relationship category. Romantic partners or spouses were the most likely to re-spond (40.8%), followed by strangers (33.3%), relatives (28.6%), coworkers (28.2%), best friends (26.2%), friends (22.9%), “other” (13.8%), and acquaintances (12.6%). The relationship type of the conversational partner is important so that the spread of information within and across social networks can be determined. People with close ties (such as best friends, relatives, romantic partners or spouses, etc.) tend to interact within similar social networks while those with weak ties (such as more casual friends, acquaintances, and strangers) are more likely to have access to distinct social networks. To enhance the “reach” into distinct social networks a word-of-mouth marketing orga-nization would likely be pleased with the response rate of strangers (one-third for this study) but would like to see a higher response rate for acquaintances (less than 13% for this study).

How Business Outcome Metrics Were Affected By Agent DisclosureThis section will report findings about how each of the outcomes discussed above were affected by whether or not the conversational partner knew of the agent’s affiliation with the word-of-mouth marketing organization. Before that, however, we will provide some background information on how many of the CPs knew of the agent’s identity, how they learned, what information was shared about the marketing organization, and what the CPs’ perceptions were of this information.

Results will continue to be reported with campaign (Type 1) versus non-campaign-re-lated (Type 2) comparisons.

How Many Agents Disclosed Their Identity?

For all episodes (both Type 1 and Type 2), 56% of the conversational partners knew of the agent’s affiliation with the word-of-mouth marketing organization while 44% did not know.

Agents were more likely to disclose for campaign-related episodes (where agents were required to disclose their identity according to the marketing organization’s policy) than for their everyday WOM episodes that were unrelated to their affiliation with the marketing organization. See Table 1 below.

Table 1 – Percentage of Conversational Partners Who Knew Agent Affiliation

The data above were obtained from the conversational partner’s survey responses. To confirm these numbers we also ran a validity check on the data based on what the agent said they disclosed versus what the CPs reported the agent did.

In about 77% of the cases the CPs and agents provided consistent responses about whether the agent did or did not disclose. In about 17% of the cases, however, agents

According to a receiver-oriented view of disclo-sure, we considered an agent to have disclosed when their conversational partners knew of their affiliation to the word-of-mouth market-ing organization through some action taken by the agent.

Type 1 (Campaign) Type 2 (Everyday) CP Knew of Agent Affiliation 63% 47%

CP Didn’t Know 37% 53%

Prior to the enforcement of the word-of-mouth marketing organizationʼs disclosure policy, 37% of the conversational partners reported they did not know of the agentʼs affiliation.

11

said they disclosed their affiliation but CPs said the agent did not (in an additional 5% of the cases, the agent said they did not disclose but the CP said they did). When comparing Type 1 and Type 2 discrepancies, the discrepancy was slightly smaller for campaign-related episodes (15.1%) than everyday episodes (17.9%).

Discrepancies in interpersonal perception are normal and to be expected due to people’s ability to recall information and different ways people make sense of their world and what is important to them14. To put this disclosure discrepancy finding in context, we report discrepancy percentages for other variables: for sex of the partici-pants there was 0% discrepancy meaning the agents and CPs were in 100% agreement about whether the other person was a male or female; for what medium was used in the WOM episode (F2F, e-mail, phone, etc.) there was 3% discrepancy; for whether or not a reference was made to another media form in the conversation (such as TV, print, internet, etc.) there was about 18% discrepancy; and for whether or not one of them was considering making a purchase for the talked-about product there was 25% discrepancy.

There could be various reasons for this discrepancy about disclosure, including 1) the agent forgot they did not disclose and said they did, 2) the agent did disclose but the CP did not remember it, 3) human data entry error when completing the web-based survey, 4) the CP lied about the agent not disclosing, or 5) the agent lied about whether or not they disclosed. One and two are somewhat plausible given the possibil-ity of forgetfulness or differences in what information people deem valuable. Human data entry error is a possibility but these were radio (single-select) buttons which are less prone to errors than drop-down menu items, for example. It is not clear what the motivation would be for the CP to lie about the agent not disclosing, though this is also plausible. Another plausible explanation is that the agent said they disclosed but did not (there is a motivation to lie since the agents took the survey via the marketing organization’s website and the current policy requires disclosure).

Are Certain Types of Relationships More Likely To Know Agentʼs Identity Rather Than Others?

There were expected, but statistically significant, differences in which relationship types were more likely to know of the agent’s affiliation. Romantic partners or spouses were the most likely to know of the agent’s affiliation (79.3%), followed by best friends (66.6%), relatives (62.3%), friends (49.2%), coworkers (36.4%), acquaintances (36%), and strangers (22%). These percentages varied slightly for Type 1 versus Type 2 episodes but there were no significant differences.

How Did the Conversational Partner Learn of Agentʼs Identity?

When the CP was aware of the agent’s identity, the CP learned by the agent directly telling them without the CP asking 75% of the time. The CP asked the agent in 7% of the interactions, it came up mutually in 6%, in 3-4% of the cases the affilia-tion came up but it was not clear who brought it up, and in less than 2% of the cases the CPs figured it out on their own. Other ways of learning about the agent identity accounted for 7-8% of the episodes, and these involved such things as the CP being present when a mailing came from the marketing organization or a minor (13-17 years of age) asking for their parent’s permission to voluntarily affiliate with the marketing organization.

There were no differences between Type 1 (campaign-related) and Type 2 (everyday) episodes regarding the frequency of unprompted and explicit disclosure. There were,

In about 15% of the campaign-related episodes there was a discrepancy between the agent and the conversa-tional partner regarding whether or not the agent disclosed.

When company affiliation was disclosed, the agent volunteered the infor-mation directly to their conversational partner without being prompted 75% of the time.

12

however, significant differences based on episode type for other disclosure strategies. First, the agent’s identity was more than four times as likely to come up mutually in Type 2 episodes. Second, the CP was almost four times as likely to ask the agent about whether or not they were affiliated with a word-of-mouth marketing organization in Type 1 episodes. These findings may suggest that conversational partners had a rea-son to think the agent was part of a word-of-mouth marketing campaign for Type 1 episodes, which then may have given rise to the CP asking the agent about their affili-ation.

What Type of Information Did The Agents Share?

We created different categories of information relevant to the marketing organization’s business model that the agent might have shared in the course of a WOM episode. We then asked CPs to report which information they knew from talking with the agent (either before or during the episode on which they reported). CPs were most likely to know that the agents receive free samples (90% knew) and that agents share their opinions with others (85% knew). Only about two-thirds knew that agents were volunteers and not paid employees of the marketing organization. Eighty-percent knew that agents report back to the marketing organization about the WOM episode, though only 45% knew the marketing organization compiles those reports for the client company who makes the brand, product, or service. Seventy-five percent knew that agents receive points for their WOM and reporting activity (67% knew these points could be redeemed for prizes, but only 34% knew the points could be donated to charity, and only 25% knew that the agent can choose not to redeem their points at all). Just over two-thirds of the CPs knew that agents received information about the product in a guide book.

What Are Conversational Partnersʼ Feelings Towards Agent Affiliation With The Marketing Company?

We wanted to find out if agent affiliation had any effect on the CP’s feelings towards the agent, the WOM episode, the brand, product, or service being discussed, or the company that makes the brand, product, or service. We distinguished the responses based on whether or not the agent had disclosed their identity in this or a prior in-teraction, or if the CP had learned of the agent’s affiliation as a result of taking this survey. Results are reported in Table 2.

Table 2 – Conversational Partners Feelings About Agent Affiliation

When CP Already Knew of Agent Affiliation No Effect Positive Negative

Feelings toward agent* 78.8% 21.2% 0%…Toward episode 74.6% 22.9% 2.5%…Toward OBPS* 70.3% 29.7% 0%

65.3% 33.0% 1.7%When CP Did Not Know of Agent Affiliation

No Effect Positive Negative

77.4% 15.1% 7.5%

75.3% 19.4% 5.4%73.1% 20.4% 6.5%

77.4% 18.3% 4.3%

…Toward company*

Feelings toward agent*…Toward episode…Toward OBPS*…Toward company*

* Differences are significant at p < .05

13

Overall, knowing that a person was affiliated with a marketing organization had no effect or a positive effect to the CPs’ feelings about the agent, the WOM episode itself, to the brand, product, or service being discussed, or towards the company. The “no ef-fect” percentages were remarkably consistent regardless of whether or not the CP knew of the agent’s affiliation or not, or based on the type of relationship between the CP and agent (stranger, friend, relative, etc.). Thus, for about 75% of the CPs in this study the fact that the person with whom they were talking was affiliated with a word-of-mouth marketing organization was a “non-issue.”

There is, however, a noticeable increase in negative feelings when the CP did not al-ready know about the agent’s affiliation. This indicates a “backlash” effect for approxi-mately 5% of the CPs. The differences are statistically significant for feelings towards the agent, the product/service, and the company that makes the product/service.

When we look at just those CPs with negative feelings toward the agent, the WOM episode, the product or service, or the company that makes the product or service, we find that those with negative feelings are more likely to be strangers, acquaintances, friends, or coworkers (rather than best friends, relatives, or romantic partners/spouses).

To provide a richer sense of CPs feelings about agent affiliation, and differences be-tween whether or not the agent had disclosed to the CP, we include a few qualitative comments from the surveys (gray box on following page). The quotations are presented according to whether the CP knew of the agent’s affiliation or not and whether their feelings were positive, negative, or neutral. Comments were only edited for ease of reading by fixing typos and punctuation. All quotations are from campaign-related (Type 1) episodes unless otherwise indicated.

Does Knowing Agent Affiliation Matter To Outcomes?

Whether or not the CP knew of the agent’s affiliation did not matter towards any of the key business outcomes except one. That is, disclosure did not affect inquiry likelihood or behavior, use or purchase likelihood or behavior, or pass-along/relay likelihood.

However, disclosure did affect how many people a conversational participant told about the brand, product, or service. For episodes where CPs knew they were talking with a word-of-mouth marketing agent they told, on average, just over two people (average = 2.02), whereas when the CPs did not know they were talking with an agent they told, on average, just over one other person (average = 1.18). Thus, the pass-along/relay rate was 70% higher when CPs were aware of the agent’s participation in the organized word-of-mouth marketing program. The statistical significance of this 70% increase is sensitive to how “outlier” cases are handled. In about 8% of the cases CPs reported telling five or more people about the OBPS (and in one case, a CP told 20 people). Although these numbers are all very plausible, the outlier cases skewed the pass-along rates higher and, for whatever reasons, were far more likely to occur when the CP knew of the agent’s affiliation and when the brand, product, or service being discussed was part of an organized campaign (campaign-related episodes had a higher pass-along/relay rate than everyday episodes – averages of 1.87 versus 1.36, respectively – but this difference in itself was not statistically significant).

It is not immediately clear why the pass-along/relay rate is higher for episodes where the agent’s affiliation was known. It could be that CPs who know they are talking with a participant in an organized word-of-mouth marketing program perceive they are getting new, or inside, information and thus are more likely to pass-along/relay that information to others. Additionally, it could be related to ratings of conversational

For about 75% of the conversational partners in this study the fact that the person with whom they were talking was affiliated with a word-of-mouth marketing organization was a “non-issue.”

For about 5% of the conversational partners who were not aware of the agentʼs affiliation with the marketing organization there was a negative “backlash” effect when they found out.

When agents disclosed the affiliation to their conversational partner, the pass-along/relay rate was 70% higher than for those episodes where the conversational partner did not know about their affiliation.

Both Type 1 and Type 2 WOM episodes were rated as more relaxed, person-al, formal, and in-depth when the conversational partner was aware of the agentʼs affiliation.

14

Conversational Partners Knew Agentʼs Affiliation“I feel much more positive knowing she can try these products first without buying, and pass it on to me.”

– 50-year old female relative, known agent for 26 years, strong positive feelings towards product/service and company

“I know this person would not recommend anything that doesn’t work well, so I am confident in her opinions. It’s great to be able to try things out, and being able to turn your friends onto cool products is cool.”

– 25-year old female, known agent for five years, all strong positive feelings (Type 2 episode)

“I am a close friend of hers so I have always trusted her opinion, regardless of her affiliation with a word of mouth orga-nization. If she asks me to try something she enjoys, I would try, and see if I like it as well.”

– 39-year old female friend, known agent for 12 years, no strong feelings either way

“Well she is my friend and she is a good one so I don’t doubt that. But if she was to pressure me into purchasing a prod-uct or to bring up a product she thinks I should try out, then I would become skeptical.”

– 18-year old female friend, known agent for 1 year, no strong feelings either way

“When someone is involved in a marketing agency, I am suspicious of how much they actually like the product, and how much they are just trying to ‘sell’ me.”

– 22-year old female friend, known agent for 14 years, slightly negative feelings towards their word-of-mouth episode

• • •

Conversational Partners Did Not Know Agentʼs Affiliation“She has an in-depth knowledge of a product she is talking about. Like she’s done research and has a person’s best interest at heart.”

– 37-year old female relative, known agent all her life, all strong positive feelings

“Well it makes me think she is more in the know because she tried the jeans and has more info on the company and the jeans than I would get from a person in the store who just wants me to spend money.”

– 19-year old female best friend, known agent all her life, slightly positive feelings towards word-of-mouth episode and product/service

“Well she always tells me of new things she tries so if she can get credit for it, more power to her.”– 48-year old female relative, known agent 42 years, no strong feelings either way

“It doesn’t change my opinion in any way. The most important thing for me is that I get honest opinions about a prod-uct or service when I ask about it or when a person talks to me about it. I feel that most people really tell you what they think of something.”

– 63-year old female relative, known agent for 39 years, no strong feelings either way (Type 2 episode)

“At first when I read the information about the marketing agency I was shocked but after thinking about the conversa-tion again I see that it was just conversation. And I wasn’t given a commercial so I think it was okay. Seem[s] like a dif-ferent way to advertise but then it wasn’t really advertising.”

– 50-year old female friend, known agent for 5 years, slightly negative feelings towards word-of-mouth episode, product/service, and company

“I kind of feel like she was looking for a reason to tell me about this store. It doesn’t seem as natural now that I know she had some sort of hidden agenda. When I mentioned that to her she explained that it wasn’t the case at all and that it kind of fell into place where she realized she was having a conversation she could use for this study. But it still made me feel a little funny, like it wasn’t quite as sincere as I would have thought without getting this card to take part in this study.”

– 33-year old female subordinate, known agent/boss for a year, slightly negative feelings towards agent and WOM-episode (Type 2 episode)

“She was very sure and confident. I am sure she was honest of her feelings but am disappointed she did not tell me she does this all the time.”

– 51-year old female stranger, just met agent, slightly negative feelings towards agent and product/service (Type 2 episode)

15

quality and perceptions of trustworthiness and sincerity. CPs reported that both Type 1 and Type 2 episodes were of a higher conversational quality (specifically, more re-laxed, personal, formal, and in-depth) when they knew the status of the agent. Further, CPs rated agents who disclosed as more trustworthy and more caring than those agents who did not disclose their affiliation. Additional analyses revealed that perceptions of trustworthiness and caring/goodwill as well as conversational quality variables have some, albeit small, correlation with pass-along/relay likelihood.

Conclusion and Recommendations1. Participation in an organized word-of-mouth marketing program does

not undermine the effectiveness of word-of-mouth communication. This study offers evidence that the person’s participation in the organized word-of-mouth marketing program is a non-issue or a positive connota-tion for two-thirds to three-fourths of the conversational partners (i.e., non-agents) that participated in this study. This study also showed that all key outcome metrics for word-of-mouth marketing organizations were consistent whether or not the conversational partner knew of the agent’s affiliation. As in everyday, “organic” word-of-mouth episodes, where no corporate affiliation is relevant, what seems to matter is that one party perceives the other as trustworthy (being perceived as genuine and ethi-cal), caring/goodwill (having the other’s best interests at heart), and that their contributions to the conversation are relevant and valuable to the other person’s everyday life. If corporate affiliation can be used to increase trustworthiness and caring towards the other, as was shown in this study by honest disclosures, then participating in an organized campaign affiliated with a word-of-mouth marketing organization can actually be a positive.

2. Disclosure has practical business benefits. There are at least two key benefits of disclosure made clear by the results of this study. First, pass-along/relay rates can actually increase when agents disclose their affilia-tions (meaning that in an episode where X tells Y about a product, Y will, on average, pass-along/relay that information to more people). While the underlying dynamics for this are not yet fully understood they likely have to do with greater feelings of goodwill and trustworthiness, or the perception that the conversational partner is getting new information to which they would otherwise not have access. Second, the negative backlash associated with non-disclosure can be avoided and it is in the business interests of the word-of-mouth marketing organization to do so. Even though an agent’s af-filiation with a word-of-mouth marketing organization was a non-issue for a large majority of conversational partners, some CPs who were not aware of the agent’s affiliation had clear negative feelings towards the agent, the product/service being talked about, and the company that made the prod-uct/service. However, when the CP was aware of the agent’s identity there were virtually no negative feelings generated by knowledge of the agent’s affiliation. If disclosure is not present and the conversational partner feels like the agent is being “sneaky” or “deceptive” then the CP may not only walk away feeling negative towards the conversation itself but the negative feelings can be transferred to the brand and company that is paying for the services provided by the word-of-mouth marketing organization.

16

3. Disclosure does not necessarily interrupt the “natural” flow of conver-sation. Based on the results from this study, when the conversational part-ner was aware of the agent’s identity, the CP learned about it through ex-plicit, unprompted disclosure by the agent in more than 75% of episodes. Further, there were no significant differences in conversational quality or key outcome metrics when this explicit strategy was used in comparison to other strategies (such as the partner asking or the affiliation coming up mutually). Brands, products, and services are part of our everyday lives and conversations. Existing research has found that, on average, such conversa-tions take place in anywhere from 13% to 18% of conversations for people not affiliated with a word-of-mouth marketing organization (and more for those who are affiliated)15. Many word-of-mouth marketing agents simply incorporate their affiliation with the marketing organization as part of their existing brand-related conversations as they carry on relationships with friends, family members, co-workers, strangers and acquaintances.

4. Clear, enforceable disclosure policies are essential for word-of-mouth marketing organizations. Although a clear majority of agents did disclose their affiliation in campaign-related episodes (63%), consistent with the marketing organization’s policy requiring agents to disclose, 37% of the conversational partners did not know that the person with whom they were talking was affiliated with a word-of-mouth marketing organization. In part as a result of this study BzzAgent, Inc. has instituted additional pro-cedures to ensure that agents are disclosing their identities16. For example, agents are now required to indicate they have disclosed their affiliation to their conversational partners on every report they file back to BzzAgent. If agents fail to do this a certain number of times then they are required to participate in an additional training session on disclosure. Further, until they satisfactorily complete the training they are restricted from filing ad-ditional reports and are not allowed to participate in any other campaigns. Other marketing organizations that seek to organize program participants in campaigns should consider similar policies (also see point #10 below).

5. Education should be one strategy for ensuring compliance with disclo-sure policies… People affiliated with word-of-mouth marketing organiza-tions should be educated about the ethics of disclosure as well as the practi-cal business benefits. The view that explicit, unprompted disclosures are somehow “unnatural” should also be exposed as a myth. Making disclosure a part of the marketing organization’s culture should help to increase the disclosure rate. Further, agents should be especially vigilant about ensuring strangers and acquaintances are aware of their affiliation because 1) these relationship types were the least likely to know about the agent’s affilia-tion, and 2) strangers and acquaintances were more likely to have negative feelings towards the agent, interaction, brand, or company when they later find out about the marketing organization affiliation. The negative feel-ings among these relationship types are likely because these conversational partners have limited relationship history with the agent that would allow them to interpret their actions in a more sincere and trustworthy manner.

6. Enforcement should be another. Setting aside the possible explanations of human error and forgetfulness, about 15% of the agents in campaign- related episodes said they disclosed but their conversational partners did

Marketing companies should be especially vigilant about ensuring that agents make strangers and acquaintances aware of their participation in the organized word-of-mouth marketing program.

17

not validate that information. Adding a system where agents have to indicate the conversational partner was aware of their affiliation with each report should increase the disclosure rate because it raises agents’ awareness, but might also have the unintended consequence of agents simply checking the box because they know they will have to go through a training ses-sion if they do not do it, thus zeroing out any gains. If this study’s results are any indication, inaccurate information about agent disclosure might continue to range from 15% – 20%. Agencies should be especially vigilant about relationship types such as strangers, acquaintances, and co-workers as CPs in these relationships were the least likely to know the agent’s identity. These relationship types could be flagged for extra attention in enforcing disclosure policies.

7. Does disclosing agent affiliation go far enough? The results of this study raise the issue of what information needs to be disclosed. For example, is it enough just to say that an agent is affiliated with a marketing organization, or does the agent also have to disclose that they receive points for their participation in the program, that they report back to the marketing orga-nization about the WOM episode, and that these reports are summarized and presented to a client company (thus engaging in market research)? The recommendation from this study is that agents should be required to disclose their affiliation with the marketing organization and also required to not withhold any of the information about how the campaign process works. Further, WOM marketing organizations should clearly indicate on their websites how a campaign works, and for those organizations that col-lect reports back from their program participants, the website explanation should clearly indicate the reporting process and that these reports repre-sent a form of market research (only 45% of the conversational partners knew that reports were compiled and presented to a client company as part of a market research report). As a marketing organization it would make sense to encourage agents to discuss how the process works for interested conversational partners as this might only increase the depth and quality of understanding in the interaction. One conversational partner actually viewed the market research aspect of this marketing organization’s business model as a positive sign because she perceived that her agent contact had a way to provide feedback to the company (see the endnote17 for this CP’s rich explanation).

8. Do brands, products, and services from past campaigns fall under the disclosure policy? This study made a distinction between campaign-re-lated (Type 1) WOM and non-campaign-related (Type 2) WOM episodes. There seems to be no reason why an agent would need to disclose their affiliation for non-campaign-related WOM episodes since they are not receiving rewards from the marketing organization for talking or reporting about these brands. But what about brands from past campaigns? Most agents (86%) categorized WOM episodes that involved a brand from a past campaign as Type 1 (or campaign-related episode) but some (14%) considered them everyday WOM episodes18. This becomes a tricky issue because the agent does not receive rewards for talking about products from past campaigns nor do they report back to the marketing organization about them. At the same time, the agent might not otherwise talk about a

Marketing companies should clearly indicate on their website how a campaign works, and for those companies that collect reports back from their program participants, the website explanation should clearly indicate the reporting process and that these reports represent a form of market research.

To determine whether agents should disclose in episodes involving brands from past campaigns, agents should apply the three principles of trust-worthiness, goodwill, and relevance that are the bedrock of all WOM communication. If any of these are undermined by not disclosing then the agent should disclose.

18

particular product if they were not at one time involved in the campaign or aware that the campaign was going on with other agents. Since the agent is not receiving any rewards for spreading the brand commentary and because there is no reporting back to the marketing organization then it would seem unnecessary for the agent to disclose their affiliation for episodes in-volving past campaigns. As in all cases, however, this study’s findings would suggest that agents apply the three principles of trustworthiness, goodwill, and relevance that are the bedrock of all WOM communication. If any of these principles are in jeopardy of being undermined by not disclosing, then agents should choose disclosure.

9. What about the disclosure issues with word-of-mouth marketing or-ganizations and minors? Based on reports from conversational partners, word-of-mouth episodes involving minors (13-17) as agents accounted for less than 6% of the participants in this study. All of the conversational partners in the study were over 18 years of age. There were no significant differences in terms of disclosure rates for minors versus other age groups, nor in the outcome variables. One difference was that one of the conver-sational partners learned of the agent’s identity when the agent, who was a minor, asked their parent (the CP) for permission to participate in the program. This report would recommend that the same rules that apply to other program participants regarding disclosure also apply to minors, with the added element that minors should also disclose their participation to their parents and gain their permission before affiliating themselves with a word-of-mouth marketing organization.

10. Do these findings and recommendations just apply to individuals af-filiated with word-of-mouth marketing organizations? Increasingly a range of word-of-mouth marketing organizations are developing custom-branded solutions for an individual company’s brand advocates and evan-gelists, selected from the company’s own customer relationship databases. Participants should also disclose their involvement in these programs when engaging in activities related to any organized word-of-mouth program where they are receiving any type of reward or reporting back to the company on their activities. Thus any marketing organization or company that enlists the support of consumers to spread word-of-mouth as part of an organized program or campaign can revisit their business practices and disclosure policies in light of the findings and conclusions of this report.

For additional details about this summary report, please contact the study’s principal investigator, Dr. Walter Carl.

Contact Information Walter J. Carl, Ph.D.

Assistant Professor Department of Communication Studies Northeastern University, Boston, MA 02115

Phone: 617.373.4075 E-mail: w.carl@neu.edu Faculty Web Site: http://www.waltercarl.neu.edu Research Blog: http://www.wom-study.blogspot.com

Dr. Walter Carl is an Assistant Professor in the Department of Communication Studies at Northeastern University in Boston, MA. He is also an Advisory Board member of the Word-of-Mouth Marketing Association and an active participant on their Research & Metrics Council. For more infor-mation about WOMMA see http://www.womma.org.

19

Copyright and Citation Information“To Tell Or Not To Tell?” summary report and the word-of-mouth image is copyright © 2006 by Walter J. Carl, Ph.D.

Visit http://www.waltercarl.neu.edu/downloads to download additional copies of this report or to view the summary report online.

Feel free to comment and discuss in online and offline environments but please do not post the electronic version of this summary report on your server. Instead include a link to the URL above and reference the main title of the report “To Tell Or Not To Tell?”

Permission to reproduce and/or distribute “To Tell Or Not To Tell?” for commercial purposes must be requested in writing and sent to the following address: w.carl@neu.edu

Appendix A – GlossaryAgent – A generic term used to refer to any person affiliated with a marketing organi-zation and whose affiliation involves them spreading commentary and recommenda-tions about a company’s brand, product, or service. In reporting results from this study agents refer to people voluntarily affiliated with BzzAgent, Inc. (or “BzzAgents”).

CP – Conversational Partner. These were the people with whom the word-of-mouth marketing agents engaged in a WOM episode. Except in the rare cases where an agent reported talking with someone else who happened to be affiliated with the market-ing organization, conversational partners were not affiliated with the word-of-mouth marketing organization.

Disclosure – From a “receiver-oriented” view of communication, disclosure involves the conversational partners’ knowledge of the agent’s affiliation with the word-of-mouth marketing organization through some action taken by the agent.

OBPS – An acronym for organization/company, brand, product, or service. The object of most campaign-related WOM episodes are brands or products, but non-cam-paign-related WOM episodes could be an organization or company, a brand, a prod-uct, or a service.

Type 1 WOM – Refers to WOM episodes that include talk about a brand that is part of an organized word-of-mouth marketing campaign. Also known as “institutional,” “amplified,” or “campaign-related” WOM.

Type 2 WOM – Refers to WOM episodes that include talk about any brand that is not part of an organized word-of-mouth marketing campaign. Also known as “every-day,” “organic,” or “non-campaign-related” WOM.

Word-of-Mouth (WOM) Episode – Defined as any interaction, or portion of a larger interaction, that includes informal, evaluative (positive or negative) talk about an organization, brand, product, or service, which may or may not include a recommen-dation.

20

Appendix B – Methodology and LimitationsWe gave agents two sets of cards to distribute to their conversational partners, one set to be used for Type 1 (campaign-related) episodes and the other set for Type 2 (everyday WOM about non-campaign-related objects). The front of the card was es-sentially a business card for the principal investigator indicating contact information and university affiliation. The back of the card had a unique conversation code and the URL for the web-based survey. The cards made no reference to the word-of-mouth marketing organization. The marketing organization’s information was not included on the card since one of the research questions was whether or not the agents disclosed their affiliation. By not including the marketing organization’s information on the card it would be up to the agent to indicate that the study was being conducted in partner-ship with the word-of-mouth marketing organization. Consistent with the marketing organization’s procedure for campaign participation, agents received a copy of the marketing organization’s disclosure policy requiring them to disclose their affiliation to each conversational partner.

Agents were instructed to give out the cards to conversational partners when they engaged in their normal word-of-mouth activity (agents were given cards for up to six conversational partners). In the event the WOM episode took place over the phone or online they were instructed to provide the information on the card to their con-versational partner. Agents could choose to which conversational partners they would give their cards, as long as a Type 1 and a Type 2 episode were represented. Agents completed their version of the survey through the marketing organization’s web site, while the conversational partners completed their version of the survey through a third-party service that specializes in web-based surveys for academic research purposes (PsychData.com). Conversational partners were required to read and provide informed consent that they were taking part in an academic research project. The consent form purposefully did not indicate that the study was being conducted in partnership with a word-of-mouth marketing organization so that we could determine whether disclo-sure had taken place. After answering survey questions about particulars of the WOM episode (such as when and where it took place, topic and stimulants for the episode, quality of the interaction, valence of the talk, whether or not a recommendation or referral was made, etc.) a question was then asked about whether they were aware of their conversational partner’s (i.e., the agent) affiliation with a word-of-mouth market-ing organization. CPs received a different set of questions depending on their answer to this question. At the end of the survey the CPs were informed that this project was conducted in partnership with the word-of-mouth marketing organization. CPs were offered the chance to participate in random drawings for a $25 electronic gift certifi-cate to Amazon.com or Barnes & Noble. The study was approved by the Division of Research Integrity at Northeastern University.

The conversion rate for the study (the number of agent surveys per the number of sur-veys completed by conversational partners) was 27.7% (agents completed 956 surveys while CPs completed 265 surveys). Of these, 833 agent surveys and 211 CP surveys were valid and complete. The response rate for valid and complete surveys was 25.3%.

There are some limitations to this study that should be noted. First, we would like to see a higher and more even representation of relationship types, especially by strang-ers and acquaintances, as these latter two relationship types were under-represented in the study. This would enable more rigorous comparisons among different relationship categories. It is not surprising, however, that strangers and acquaintances were not as

21

well represented since a) existing research has shown that of all the relationship types, interactions with strangers and acquaintances are least likely to include brand-related WOM episodes19, and b) the WOM marketing organization encourages agents to not artificially create conversations with people they wouldn’t otherwise talk with in order to bring up brand-related talk. Second, the negative “backlash” feelings associated with not knowing about agent affiliation would likely be even higher given the findings that strangers and acquaintances were the least likely to know of the agent affiliation, and more likely to have stronger negative feelings when they did find out. Third, most of the outcome data was based on self-report of the conversational partners without objective verification (for example, the CP reported that they spoke with three other people about the brand after talking with the agent but this information was not veri-fied with another methodology, such as direct observation). The study did, however, employ a dyadic design, which is a more rigorous methodology than many other stud-ies. This was especially valuable as a “check” on agents’ self-report on their disclosure activities. Further, the response rate to the survey was an objectively verifiable outcome measure that is arguably akin to inquiry behavior (that is, where a person seeks out additional information, such as going to a website). Fourth, some of the responses by the conversational partners may be subject to a social desirability effect wherein they may have provided answers that they felt they “should” be providing. There may be some evidence of this when some CPs (about 5%) reported that the agent disclosed to them but the agent did not report the disclosure (for example, perhaps the CP felt they should have answered that the agent disclosed so that the agent would not get in “trouble”). In the very few cases this took place, the CP-agent relationship was an acquaintance, friend, or relative so the existing relational connection may have com-pelled them to respond in this way. Fifth, in order to provide a more useful point of reference the key business outcomes reported in this study should be compared with the same outcome measures from everyday word-of-mouth episodes where no party has a corporate affiliation. This way we can determine whether or not these outcome measures are stronger or weaker than everyday, organic word-of-mouth.

AcknowledgementsI would like to thank the following people for their contributions to this research project:

From BzzAgent: Matt McGlinn, Tim Matthews, Dave Balter, Jon O’Toole, Greg Bowdne, Brady Bonus, Karmen Chong, and Seth Wylie.

From Northeastern University: Ben Evarts and Dr. Carey Noland.

A special thank you goes to all of the agents and conversational partners who participated in this research project.

Funding for this project was provided in part by a Research & Scholarship Development Fund Grant from the Provost’s Office at Northeastern University, as well as by BzzAgent, Inc.

22

Notes and References(Endnotes)

1 Academic journal publication will constitute official reporting of results from this study.

2 Numerous academic and industry studies cite the effectiveness of word-of-mouth communication and marketing. For a useful overview of academic work and reasons why word-of-mouth has these effects see Francis A. Buttle’s 1998 article in Journal of Strategic Marketing, 6, 241-254 entitled “Word of Mouth: Understanding and Managing Referral Marketing.”

3 Is Buzz Marketing Illegal” in Advertising Age, October 4, 2005, the consumer watchdog group Commercial Alert’s letter to the Federal Trade Commission regarding buzz marketing at http://www.commercialalert.org/buzzmarketing.pdf, and the Word-of-Mouth Marketing Association’s (WOMMA) response to both the Advertising Age article and Commercial Alert letter at http://womma.org/pages/2005/10/womma_opposes_s.htm

4 See WOMMA’s elaboration on unethical marketing practices at http://www.womma.com/wom101f.htm

5 See quote by John Maron, Sony Ericcson’s marketing director, in the 60 Minutes special “Undercover Marketing Uncovered” at http://www.cbsnews.com/stories/2003/10/23/60minutes/main579657.shtml

6 For more details about how a BzzAgent, Inc. campaign works see http://www.bzzagent.com/pages/howitworks/howitworks.jsp

7 For more details about these key word-of-mouth metrics see “Mapping the Conversational Geog-raphy of Word-of-Mouth Marketing: An Application of the Word-of-Mouth Terminology Frame-work” from Measuring Word-of-Mouth, Volume 1 (pp. 61-69) by Dr. Walter Carl. Volume published by the Word of Mouth Marketing Association.

8 The WOMMA Terminology Framework can be accessed at http://www.womma.org/terminology.htm

9 WOMMA uses a similar distinction between “organic” (everyday) and “amplified” (institutional) word-of-mouth (see http://www.womma.org/wom101d.htm). For more details on the distinction between everyday and institutional word-of-mouth episodes, see “What’s All the Buzz About? Ev-eryday Communication and the Relational Basis of Word-of-Mouth and Buzz Marketing Practic-es,” Management Communication Quarterly, Volume 19, Issue 4, 2006 by Dr. Walter Carl. Pre-press version available at http://www.waltercarl.neu.edu/downloads/

10 For additional details on terminology distinction see the Word of Mouth Marketing Association’s primer on word-of-mouth at http://www.womma.org/wom101b.htm as well as the introduction to the book Connected Marketing: The Viral, Buzz, and Word of Mouth Revolution edited by Justin Kirby and Paul Marsden (Butterworth-Heinemann, 2006).

11 This description was provided towards the end of the survey after other information regarding the particulars of the WOM episode and outcomes were obtained (thus not biasing the outcome metrics) and, in the case of conversational partners who were not aware of the agent’s affiliation, to enable comparisons before and after learning this information.

12 The three credibility factors were derived from measures of source credibility in communication research. More details can be found in McCroskey, J. C., & Teve, J. J. (1999). Goodwill: A reex-amination of the construct and its measurement. Communication Monographs, 66, 90-103.

13 For details on BzzAgent’s corporate disclosure policy see http://www.bzzagent.com/value_of_disclosure.pdf

14 The most sustained and sophisticated research tradition on interpersonal perception can be found in the work of Dr. David A. Kenny and colleagues. To learn more, see his book Interpersonal Percep-tion: A Social Relations Analysis (1994) published by Guilford. To learn more about how people make sense of the world within the context of close relationship see Dr. Steve Duck’s book entitled Meaningful Relationships: Talking, Sense, and Relating (1994) by Sage Publications.

15 See “What’s All the Buzz About?” study (full citation above) and http://wom-study.blogspot.com/2005/12/how-much-we-do-word-of-mouth-marketing.html

16 For BzzAgent’s press release about the enhanced disclosure policy, see http://www.bzzagent.com/downloads/press/BzzAgent_Enhances_Disclosure_Policy.pdf

23

17 This quotation, the conversational partner’s free response about the impact of the WOM episode, was provided by a 35 year-old female whose sister is an agent: “I think it was cool that she [agent] mentioned a product I had tried and that I could tell her the cinammon [sic] flavor was not worth buying. She wasn’t forcing it into conversation because it just came up and related to what we were talking about. She is always sharing info about new products, and I actually think it is neat to know she is getting to try new products and tell me about them. Also, she has ALWAYS enjoyed telling companies what she thinks, whether good or bad. I laughed when she later told me that one of her reports mentioned wanting the bowl to have a measurement for the water level. That sounds just like something she’d do, and by the way I agree with that. And now I know I can give feedback to companies through her. Like in our conversation, I wanted her to report that I think [name of food brand] should make a brownie with peanut butter and one that has cherries. Yumm.”

18 Of the episodes identified by the agents as “Type 1” (campaign-related), 86% were from a current campaign, 13% were from a past campaign, and 1% was not part of a campaign (thus incorrectly labeled as Type 1 when it should have been labeled Type 2). Of the episodes identified by the agents as “Type 2” (non-campaign-related), 1% was from a current campaign (thus incorrectly labeled as Type 2 when it should have been labeled Type 1), 2% were from past campaigns, and 97% were not part of a campaign. When “cleaning” the data we ensured that episode types were correctly labeled, but left it up to the agent’s discretion as to whether or not a brand from a past campaign was labeled as Type 1 or Type 2.

19 See the episode-to-interaction ratio results reported in the “What’s All the Buzz About?” study (full citation above). The WOM episode-to-interaction ratio shows the percentage of all interactions (about any topic) that include brand-related talk. When this metric is broken down by relational type, best friend relationships are the most likely to include brand-related talk (around 33% of all best friend interactions include a WOM episode), followed by relatives (28%), friends (27%), romantic partners/spouses (25%), acquaintances (24%), coworkers (23%), and strangers (23%).

Visit http://www.waltercarl.neu.edu/downloads to download additional copies of this report or to view the summary report online.